### Table 1: Inputs To Fixed-point Math Analysis

"... In PAGE 4: ... An example model to illustrate fixed-point math analysis is depicted in Figure 2. The inputs to the fixed-point analysis are ranges and resolutions of the interface variables as listed in Table1 . The results of the fixed- point math analysis are summarized in Table 2.... In PAGE 7: ... Further, the input test case generation is also eased by the simulation ability of the model. Hence, the unit testing effort comparison provided by Table1 shows a considerable reduction in effort for MBSD approach. 6.... In PAGE 7: ... Our comparison studies in Table 4 were based on autocode generated for HC12 microcontroller and Cosmic complier combination. Table1 : Critical Resource Comparison Ratio between Autocode and point hand code ROM 1.00 RAM 1.... ..."

### Table 2. Fixed-point precision requirements for the RAKE receiver.

"... In PAGE 6: ... A quantization analysis tool developed at the University of Texas, Dallas [7] was used for determining the dynamic range and precision requirements of the RAKE receiver. Table2 shows the fixed-point integer requirements of the individual RAKE receiver variables after quantization analysis. From the obtained fixed-point formats, extensive simulations were carried out using MAT- LAB/C with C++ classes in SystemC providing the fixed-point arithmetic support.... ..."

### Table 1: Number of compute kernels implemented per FPGA for the double-precision and fixed-point kernels.

2007

"... In PAGE 3: ... The exact number of such simultaneous kernels depends on the size of the FPGA chip used. Table1 summarizes the number of unrolled steps implement on each of the chips of each of the MAPs available in our system. We use the MAP Series C processor to compute both autocorrelation and cross-correlation and the MAP Series E processor to ... In PAGE 5: ... We also benefit from the ability to implement an arbitrary precision numerical data type. Thus, when using double-precision floating-point for the bin boundaries, as shown in Figure 3, we can place 2 kernels per chip for the autocorrelation subroutine and 3 kernels per chip for the cross-correlation subroutine (last but one column in Table1 ). However, a more detailed numerical analysis shows that a 43-bit, fixed- point data type is sufficient to cover the necessary range of scales used to store the bin boundaries in this particular application.... In PAGE 5: ... However, a more detailed numerical analysis shows that a 43-bit, fixed- point data type is sufficient to cover the necessary range of scales used to store the bin boundaries in this particular application. Therefore, when we replace the double-precision floating-point comparison operator with a custom, 43-bit fixed-point comparison operator [14], we can place a larger number of kernels per chip (last column in Table1 ). The rest of this paper focuses on this fixed-point implementation.... ..."

Cited by 1

### Table 3: Comparison of Fixed-point Approximation and Exact Solution of Blocking Probability

2000

"... In PAGE 51: ...4.2 Approximation Table3 compares the blocking probability calculated by an exact analysis (so- lution to the multi-dimensional Markov chain) with the xed-point approximation (isolating one cell) for a simple network of two cells. We do not compare the approxi- mation of isolating two cells here because for this network, the two-cell approximation would be exact.... ..."

### Table 1: Fixed points and stability conditions for com- petitive dynamics

"... In PAGE 3: ... Because the stability of each equilib- rium point changes depending upon the values of the parameters i and ji, we also computed a set of sta- bility conditions. The results of the analysis are shown in Table1 , revealing eight unique equilibrium points, corresponding to eight unique behaviors that can be generated by the agent. The stability analysis reveals three classes of stable solutions.... ..."

### Table 2: Fixed Point Distribution for the uni100m Data

2006

"... In PAGE 5: .... QK, the number of queries that had to search K cells to find the closest point, for various K. Since the numbers cover such a wide range, tables present them better than graphs. Table2 shows CK for the uni100m dataset with Nf = 100, 000, 000 and G = 232. Table 3 shows QK.... ..."

### Table 1: An Example of Fixed-Point Addition

"... In PAGE 5: ... Detailed implementation of the Fixed-Point adder based on a Ripple Carry Adder and a Carry Lookahead architecture are shown in Figure 12 and Figure 13 respectively. Table1 gives an exam- ple of this parameterized adder. The term parameterized indicates that the length of integer part and fractional part can be set through two generic variables: int lens and frac lens for the length of integer part and fractional part separately.... ..."

### Table 1: Construction of the fixed-point

2001

"... In PAGE 3: ...Table1... ..."

Cited by 12

### TABLE VIII Fixed-point Precision and Compliance

### Table 2: Fixed-point DSPs.

"... In PAGE 2: ... The control from a microprocessor can be achieved via a host in- terface with a word-length of 8 bits. An overview of xed-point DSPs with respect to word-length and cycle time is shown in Table2 . Ba- sically, the precision of the arithmetic can be dou- bled if quantization a ects the stability and numeric precision of the applied algorithm.... ..."